vendor/llvm-project/llvmorg-14-init-10186-gff7f2cfa959b

21 files changed, 551 insertions, 304 deletions

diff --git a/llvm/lib/CodeGen/AsmPrinter/ARMException.cpp b/llvm/lib/CodeGen/AsmPrinter/ARMException.cpp
index db4215e92d44..223840c21d8b 100644
--- a/llvm/lib/CodeGen/AsmPrinter/ARMException.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/ARMException.cpp
@@ -75,7 +75,6 @@ void ARMException::endFunction(const MachineFunction *MF) {
     // Emit references to personality.
     if (Per) {
       MCSymbol *PerSym = Asm->getSymbol(Per);
-      Asm->OutStreamer->emitSymbolAttribute(PerSym, MCSA_Global);
       ATS.emitPersonality(PerSym);
     }
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp b/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
index e528d33b5f8c..cc848d28a9a7 100644
--- a/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
@@ -71,7 +71,6 @@
 #include "llvm/IR/GCStrategy.h"
 #include "llvm/IR/GlobalAlias.h"
 #include "llvm/IR/GlobalIFunc.h"
-#include "llvm/IR/GlobalIndirectSymbol.h"
 #include "llvm/IR/GlobalObject.h"
 #include "llvm/IR/GlobalValue.h"
 #include "llvm/IR/GlobalVariable.h"
@@ -102,6 +101,7 @@
 #include "llvm/MC/MCTargetOptions.h"
 #include "llvm/MC/MCValue.h"
 #include "llvm/MC/SectionKind.h"
+#include "llvm/MC/TargetRegistry.h"
 #include "llvm/Pass.h"
 #include "llvm/Remarks/Remark.h"
 #include "llvm/Remarks/RemarkFormat.h"
@@ -115,7 +115,6 @@
 #include "llvm/Support/Format.h"
 #include "llvm/Support/MathExtras.h"
 #include "llvm/Support/Path.h"
-#include "llvm/Support/TargetRegistry.h"
 #include "llvm/Support/Timer.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Target/TargetLoweringObjectFile.h"
@@ -275,7 +274,7 @@ bool AsmPrinter::doInitialization(Module &M) {
   const_cast<TargetLoweringObjectFile &>(getObjFileLowering())
       .getModuleMetadata(M);
 
-  OutStreamer->InitSections(false);
+  OutStreamer->initSections(false, *TM.getMCSubtargetInfo());
 
   if (DisableDebugInfoPrinting)
     MMI->setDebugInfoAvailability(false);
@@ -326,16 +325,10 @@ bool AsmPrinter::doInitialization(Module &M) {
 
   // Emit module-level inline asm if it exists.
   if (!M.getModuleInlineAsm().empty()) {
-    // We're at the module level. Construct MCSubtarget from the default CPU
-    // and target triple.
-    std::unique_ptr<MCSubtargetInfo> STI(TM.getTarget().createMCSubtargetInfo(
-        TM.getTargetTriple().str(), TM.getTargetCPU(),
-        TM.getTargetFeatureString()));
-    assert(STI && "Unable to create subtarget info");
     OutStreamer->AddComment("Start of file scope inline assembly");
     OutStreamer->AddBlankLine();
-    emitInlineAsm(M.getModuleInlineAsm() + "\n",
-                  OutContext.getSubtargetCopy(*STI), TM.Options.MCOptions);
+    emitInlineAsm(M.getModuleInlineAsm() + "\n", *TM.getMCSubtargetInfo(),
+                  TM.Options.MCOptions);
     OutStreamer->AddComment("End of file scope inline assembly");
     OutStreamer->AddBlankLine();
   }
@@ -1422,7 +1415,7 @@ void AsmPrinter::emitFunctionBody() {
       });
       R << "BasicBlock: " << ore::NV("BasicBlock", MBB.getName()) << "\n";
       for (auto &KV : MnemonicVec) {
-        auto Name = (Twine("INST_") + KV.first.trim()).str();
+        auto Name = (Twine("INST_") + getToken(KV.first.trim()).first).str();
         R << KV.first << ": " << ore::NV(Name, KV.second) << "\n";
       }
       ORE->emit(R);
@@ -1610,14 +1603,13 @@ void AsmPrinter::emitGlobalGOTEquivs() {
     emitGlobalVariable(GV);
 }
 
-void AsmPrinter::emitGlobalIndirectSymbol(Module &M,
-                                          const GlobalIndirectSymbol& GIS) {
-  MCSymbol *Name = getSymbol(&GIS);
-  bool IsFunction = GIS.getValueType()->isFunctionTy();
+void AsmPrinter::emitGlobalAlias(Module &M, const GlobalAlias &GA) {
+  MCSymbol *Name = getSymbol(&GA);
+  bool IsFunction = GA.getValueType()->isFunctionTy();
   // Treat bitcasts of functions as functions also. This is important at least
   // on WebAssembly where object and function addresses can't alias each other.
   if (!IsFunction)
-    if (auto *CE = dyn_cast<ConstantExpr>(GIS.getIndirectSymbol()))
+    if (auto *CE = dyn_cast<ConstantExpr>(GA.getAliasee()))
       if (CE->getOpcode() == Instruction::BitCast)
         IsFunction =
           CE->getOperand(0)->getType()->getPointerElementType()->isFunctionTy();
@@ -1627,61 +1619,80 @@ void AsmPrinter::emitGlobalIndirectSymbol(Module &M,
   // point, all the extra label is emitted, we just have to emit linkage for
   // those labels.
   if (TM.getTargetTriple().isOSBinFormatXCOFF()) {
-    assert(!isa<GlobalIFunc>(GIS) && "IFunc is not supported on AIX.");
     assert(MAI->hasVisibilityOnlyWithLinkage() &&
            "Visibility should be handled with emitLinkage() on AIX.");
-    emitLinkage(&GIS, Name);
+    emitLinkage(&GA, Name);
     // If it's a function, also emit linkage for aliases of function entry
     // point.
     if (IsFunction)
-      emitLinkage(&GIS,
-                  getObjFileLowering().getFunctionEntryPointSymbol(&GIS, TM));
+      emitLinkage(&GA,
+                  getObjFileLowering().getFunctionEntryPointSymbol(&GA, TM));
     return;
   }
 
-  if (GIS.hasExternalLinkage() || !MAI->getWeakRefDirective())
+  if (GA.hasExternalLinkage() || !MAI->getWeakRefDirective())
     OutStreamer->emitSymbolAttribute(Name, MCSA_Global);
-  else if (GIS.hasWeakLinkage() || GIS.hasLinkOnceLinkage())
+  else if (GA.hasWeakLinkage() || GA.hasLinkOnceLinkage())
     OutStreamer->emitSymbolAttribute(Name, MCSA_WeakReference);
   else
-    assert(GIS.hasLocalLinkage() && "Invalid alias or ifunc linkage");
+    assert(GA.hasLocalLinkage() && "Invalid alias linkage");
 
   // Set the symbol type to function if the alias has a function type.
   // This affects codegen when the aliasee is not a function.
   if (IsFunction)
-    OutStreamer->emitSymbolAttribute(Name, isa<GlobalIFunc>(GIS)
-                                               ? MCSA_ELF_TypeIndFunction
-                                               : MCSA_ELF_TypeFunction);
+    OutStreamer->emitSymbolAttribute(Name, MCSA_ELF_TypeFunction);
 
-  emitVisibility(Name, GIS.getVisibility());
+  emitVisibility(Name, GA.getVisibility());
 
-  const MCExpr *Expr = lowerConstant(GIS.getIndirectSymbol());
+  const MCExpr *Expr = lowerConstant(GA.getAliasee());
 
-  if (isa<GlobalAlias>(&GIS) && MAI->hasAltEntry() && isa<MCBinaryExpr>(Expr))
+  if (MAI->hasAltEntry() && isa<MCBinaryExpr>(Expr))
     OutStreamer->emitSymbolAttribute(Name, MCSA_AltEntry);
 
   // Emit the directives as assignments aka .set:
   OutStreamer->emitAssignment(Name, Expr);
-  MCSymbol *LocalAlias = getSymbolPreferLocal(GIS);
+  MCSymbol *LocalAlias = getSymbolPreferLocal(GA);
   if (LocalAlias != Name)
     OutStreamer->emitAssignment(LocalAlias, Expr);
 
-  if (auto *GA = dyn_cast<GlobalAlias>(&GIS)) {
-    // If the aliasee does not correspond to a symbol in the output, i.e. the
-    // alias is not of an object or the aliased object is private, then set the
-    // size of the alias symbol from the type of the alias. We don't do this in
-    // other situations as the alias and aliasee having differing types but same
-    // size may be intentional.
-    const GlobalObject *BaseObject = GA->getBaseObject();
-    if (MAI->hasDotTypeDotSizeDirective() && GA->getValueType()->isSized() &&
-        (!BaseObject || BaseObject->hasPrivateLinkage())) {
-      const DataLayout &DL = M.getDataLayout();
-      uint64_t Size = DL.getTypeAllocSize(GA->getValueType());
-      OutStreamer->emitELFSize(Name, MCConstantExpr::create(Size, OutContext));
-    }
+  // If the aliasee does not correspond to a symbol in the output, i.e. the
+  // alias is not of an object or the aliased object is private, then set the
+  // size of the alias symbol from the type of the alias. We don't do this in
+  // other situations as the alias and aliasee having differing types but same
+  // size may be intentional.
+  const GlobalObject *BaseObject = GA.getAliaseeObject();
+  if (MAI->hasDotTypeDotSizeDirective() && GA.getValueType()->isSized() &&
+      (!BaseObject || BaseObject->hasPrivateLinkage())) {
+    const DataLayout &DL = M.getDataLayout();
+    uint64_t Size = DL.getTypeAllocSize(GA.getValueType());
+    OutStreamer->emitELFSize(Name, MCConstantExpr::create(Size, OutContext));
   }
 }
 
+void AsmPrinter::emitGlobalIFunc(Module &M, const GlobalIFunc &GI) {
+  assert(!TM.getTargetTriple().isOSBinFormatXCOFF() &&
+         "IFunc is not supported on AIX.");
+
+  MCSymbol *Name = getSymbol(&GI);
+
+  if (GI.hasExternalLinkage() || !MAI->getWeakRefDirective())
+    OutStreamer->emitSymbolAttribute(Name, MCSA_Global);
+  else if (GI.hasWeakLinkage() || GI.hasLinkOnceLinkage())
+    OutStreamer->emitSymbolAttribute(Name, MCSA_WeakReference);
+  else
+    assert(GI.hasLocalLinkage() && "Invalid ifunc linkage");
+
+  OutStreamer->emitSymbolAttribute(Name, MCSA_ELF_TypeIndFunction);
+  emitVisibility(Name, GI.getVisibility());
+
+  // Emit the directives as assignments aka .set:
+  const MCExpr *Expr = lowerConstant(GI.getResolver());
+  OutStreamer->emitAssignment(Name, Expr);
+  MCSymbol *LocalAlias = getSymbolPreferLocal(GI);
+  if (LocalAlias != Name)
+    OutStreamer->emitAssignment(LocalAlias, Expr);
+}
+
 void AsmPrinter::emitRemarksSection(remarks::RemarkStreamer &RS) {
   if (!RS.needsSection())
     return;
@@ -1815,6 +1826,11 @@ bool AsmPrinter::doFinalization(Module &M) {
     }
   }
 
+  // This needs to happen before emitting debug information since that can end
+  // arbitrary sections.
+  if (auto *TS = OutStreamer->getTargetStreamer())
+    TS->emitConstantPools();
+
   // Finalize debug and EH information.
   for (const HandlerInfo &HI : Handlers) {
     NamedRegionTimer T(HI.TimerName, HI.TimerDescription, HI.TimerGroupName,
@@ -1857,11 +1873,11 @@ bool AsmPrinter::doFinalization(Module &M) {
       AliasStack.push_back(Cur);
     }
     for (const GlobalAlias *AncestorAlias : llvm::reverse(AliasStack))
-      emitGlobalIndirectSymbol(M, *AncestorAlias);
+      emitGlobalAlias(M, *AncestorAlias);
     AliasStack.clear();
   }
   for (const auto &IFunc : M.ifuncs())
-    emitGlobalIndirectSymbol(M, IFunc);
+    emitGlobalIFunc(M, IFunc);
 
   GCModuleInfo *MI = getAnalysisIfAvailable<GCModuleInfo>();
   assert(MI && "AsmPrinter didn't require GCModuleInfo?");
@@ -2455,9 +2471,14 @@ void AsmPrinter::emitAlignment(Align Alignment, const GlobalObject *GV) const {
   if (Alignment == Align(1))
     return; // 1-byte aligned: no need to emit alignment.
 
-  if (getCurrentSection()->getKind().isText())
-    OutStreamer->emitCodeAlignment(Alignment.value());
-  else
+  if (getCurrentSection()->getKind().isText()) {
+    const MCSubtargetInfo *STI = nullptr;
+    if (this->MF)
+      STI = &getSubtargetInfo();
+    else
+      STI = TM.getMCSubtargetInfo();
+    OutStreamer->emitCodeAlignment(Alignment.value(), STI);
+  } else
     OutStreamer->emitValueToAlignment(Alignment.value());
 }
 
@@ -2513,7 +2534,7 @@ const MCExpr *AsmPrinter::lowerConstant(const Constant *CV) {
     OS << "Unsupported expression in static initializer: ";
     CE->printAsOperand(OS, /*PrintType=*/false,
                    !MF ? nullptr : MF->getFunction().getParent());
-    report_fatal_error(OS.str());
+    report_fatal_error(Twine(OS.str()));
   }
   case Instruction::GetElementPtr: {
     // Generate a symbolic expression for the byte address
@@ -3265,21 +3286,21 @@ void AsmPrinter::emitBasicBlockStart(const MachineBasicBlock &MBB) {
   // reference the block.  It is possible that there is more than one label
   // here, because multiple LLVM BB's may have been RAUW'd to this block after
   // the references were generated.
+  const BasicBlock *BB = MBB.getBasicBlock();
   if (MBB.hasAddressTaken()) {
-    const BasicBlock *BB = MBB.getBasicBlock();
     if (isVerbose())
       OutStreamer->AddComment("Block address taken");
 
     // MBBs can have their address taken as part of CodeGen without having
     // their corresponding BB's address taken in IR
-    if (BB->hasAddressTaken())
+    if (BB && BB->hasAddressTaken())
       for (MCSymbol *Sym : MMI->getAddrLabelSymbolToEmit(BB))
         OutStreamer->emitLabel(Sym);
   }
 
   // Print some verbose block comments.
   if (isVerbose()) {
-    if (const BasicBlock *BB = MBB.getBasicBlock()) {
+    if (BB) {
       if (BB->hasName()) {
         BB->printAsOperand(OutStreamer->GetCommentOS(),
                            /*PrintType=*/false, BB->getModule());
@@ -3538,7 +3559,7 @@ void AsmPrinter::emitXRayTable() {
   // pointers. This should work for both 32-bit and 64-bit platforms.
   if (FnSledIndex) {
     OutStreamer->SwitchSection(FnSledIndex);
-    OutStreamer->emitCodeAlignment(2 * WordSizeBytes);
+    OutStreamer->emitCodeAlignment(2 * WordSizeBytes, &getSubtargetInfo());
     OutStreamer->emitSymbolValue(SledsStart, WordSizeBytes, false);
     OutStreamer->emitSymbolValue(SledsEnd, WordSizeBytes, false);
     OutStreamer->SwitchSection(PrevSection);
diff --git a/llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp b/llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp
index 4a93181f5439..ef1abc47701a 100644
--- a/llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp
@@ -30,10 +30,10 @@
 #include "llvm/MC/MCStreamer.h"
 #include "llvm/MC/MCSubtargetInfo.h"
 #include "llvm/MC/MCSymbol.h"
+#include "llvm/MC/TargetRegistry.h"
 #include "llvm/Support/ErrorHandling.h"
 #include "llvm/Support/MemoryBuffer.h"
 #include "llvm/Support/SourceMgr.h"
-#include "llvm/Support/TargetRegistry.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Target/TargetMachine.h"
 using namespace llvm;
@@ -129,13 +129,16 @@ void AsmPrinter::emitInlineAsm(StringRef Str, const MCSubtargetInfo &STI,
 }
 
 static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
-                               MachineModuleInfo *MMI, AsmPrinter *AP,
-                               uint64_t LocCookie, raw_ostream &OS) {
+                               MachineModuleInfo *MMI, const MCAsmInfo *MAI,
+                               AsmPrinter *AP, uint64_t LocCookie,
+                               raw_ostream &OS) {
   // Switch to the inline assembly variant.
   OS << "\t.intel_syntax\n\t";
 
+  int CurVariant = -1; // The number of the {.|.|.} region we are in.
   const char *LastEmitted = AsmStr; // One past the last character emitted.
   unsigned NumOperands = MI->getNumOperands();
+  int AsmPrinterVariant = 1; // X86MCAsmInfo.cpp's AsmWriterFlavorTy::Intel.
 
   while (*LastEmitted) {
     switch (*LastEmitted) {
@@ -145,8 +148,8 @@ static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
       while (*LiteralEnd && *LiteralEnd != '{' && *LiteralEnd != '|' &&
              *LiteralEnd != '}' && *LiteralEnd != '$' && *LiteralEnd != '\n')
         ++LiteralEnd;
-
-      OS.write(LastEmitted, LiteralEnd-LastEmitted);
+      if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
+        OS.write(LastEmitted, LiteralEnd - LastEmitted);
       LastEmitted = LiteralEnd;
       break;
     }
@@ -164,6 +167,27 @@ static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
       case '$':
         ++LastEmitted;  // Consume second '$' character.
         break;
+      case '(':        // $( -> same as GCC's { character.
+        ++LastEmitted; // Consume '(' character.
+        if (CurVariant != -1)
+          report_fatal_error("Nested variants found in inline asm string: '" +
+                             Twine(AsmStr) + "'");
+        CurVariant = 0; // We're in the first variant now.
+        break;
+      case '|':
+        ++LastEmitted; // Consume '|' character.
+        if (CurVariant == -1)
+          OS << '|'; // This is gcc's behavior for | outside a variant.
+        else
+          ++CurVariant; // We're in the next variant.
+        break;
+      case ')':        // $) -> same as GCC's } char.
+        ++LastEmitted; // Consume ')' character.
+        if (CurVariant == -1)
+          OS << '}'; // This is gcc's behavior for } outside a variant.
+        else
+          CurVariant = -1;
+        break;
       }
       if (Done) break;
 
@@ -176,16 +200,15 @@ static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
       // If we have ${:foo}, then this is not a real operand reference, it is a
       // "magic" string reference, just like in .td files.  Arrange to call
       // PrintSpecial.
-      if (HasCurlyBraces && LastEmitted[0] == ':') {
+      if (HasCurlyBraces && *LastEmitted == ':') {
         ++LastEmitted;
         const char *StrStart = LastEmitted;
         const char *StrEnd = strchr(StrStart, '}');
         if (!StrEnd)
           report_fatal_error("Unterminated ${:foo} operand in inline asm"
                              " string: '" + Twine(AsmStr) + "'");
-
-        std::string Val(StrStart, StrEnd);
-        AP->PrintSpecial(MI, OS, Val.c_str());
+        if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
+          AP->PrintSpecial(MI, OS, StringRef(StrStart, StrEnd - StrStart));
         LastEmitted = StrEnd+1;
         break;
       }
@@ -201,7 +224,7 @@ static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
                            Twine(AsmStr) + "'");
       LastEmitted = IDEnd;
 
-      if (Val >= NumOperands-1)
+      if (Val >= NumOperands - 1)
         report_fatal_error("Invalid $ operand number in inline asm string: '" +
                            Twine(AsmStr) + "'");
 
@@ -228,40 +251,50 @@ static void EmitMSInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
 
       // Okay, we finally have a value number.  Ask the target to print this
       // operand!
-      unsigned OpNo = InlineAsm::MIOp_FirstOperand;
-
-      bool Error = false;
+      if (CurVariant == -1 || CurVariant == AsmPrinterVariant) {
+        unsigned OpNo = InlineAsm::MIOp_FirstOperand;
 
-      // Scan to find the machine operand number for the operand.
-      for (; Val; --Val) {
-        if (OpNo >= MI->getNumOperands()) break;
-        unsigned OpFlags = MI->getOperand(OpNo).getImm();
-        OpNo += InlineAsm::getNumOperandRegisters(OpFlags) + 1;
-      }
+        bool Error = false;
 
-      // We may have a location metadata attached to the end of the
-      // instruction, and at no point should see metadata at any
-      // other point while processing. It's an error if so.
-      if (OpNo >= MI->getNumOperands() ||
-          MI->getOperand(OpNo).isMetadata()) {
-        Error = true;
-      } else {
-        unsigned OpFlags = MI->getOperand(OpNo).getImm();
-        ++OpNo;  // Skip over the ID number.
+        // Scan to find the machine operand number for the operand.
+        for (; Val; --Val) {
+          if (OpNo >= MI->getNumOperands())
+            break;
+          unsigned OpFlags = MI->getOperand(OpNo).getImm();
+          OpNo += InlineAsm::getNumOperandRegisters(OpFlags) + 1;
+        }
 
-        if (InlineAsm::isMemKind(OpFlags)) {
-          Error = AP->PrintAsmMemoryOperand(
-              MI, OpNo, Modifier[0] ? Modifier : nullptr, OS);
+        // We may have a location metadata attached to the end of the
+        // instruction, and at no point should see metadata at any
+        // other point while processing. It's an error if so.
+        if (OpNo >= MI->getNumOperands() || MI->getOperand(OpNo).isMetadata()) {
+          Error = true;
         } else {
-          Error = AP->PrintAsmOperand(MI, OpNo,
-                                      Modifier[0] ? Modifier : nullptr, OS);
+          unsigned OpFlags = MI->getOperand(OpNo).getImm();
+          ++OpNo; // Skip over the ID number.
+
+          // FIXME: Shouldn't arch-independent output template handling go into
+          // PrintAsmOperand?
+          // Labels are target independent.
+          if (MI->getOperand(OpNo).isBlockAddress()) {
+            const BlockAddress *BA = MI->getOperand(OpNo).getBlockAddress();
+            MCSymbol *Sym = AP->GetBlockAddressSymbol(BA);
+            Sym->print(OS, AP->MAI);
+            MMI->getContext().registerInlineAsmLabel(Sym);
+          } else if (InlineAsm::isMemKind(OpFlags)) {
+            Error = AP->PrintAsmMemoryOperand(
+                MI, OpNo, Modifier[0] ? Modifier : nullptr, OS);
+          } else {
+            Error = AP->PrintAsmOperand(MI, OpNo,
+                                        Modifier[0] ? Modifier : nullptr, OS);
+          }
+        }
+        if (Error) {
+          std::string msg;
+          raw_string_ostream Msg(msg);
+          Msg << "invalid operand in inline asm: '" << AsmStr << "'";
+          MMI->getModule()->getContext().emitError(LocCookie, Msg.str());
         }
-      }
-      if (Error) {
-        std::string msg;
-        raw_string_ostream Msg(msg);
-        Msg << "invalid operand in inline asm: '" << AsmStr << "'";
-        MMI->getModule()->getContext().emitError(LocCookie, Msg.str());
       }
       break;
     }
@@ -274,10 +307,10 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
                                 MachineModuleInfo *MMI, const MCAsmInfo *MAI,
                                 AsmPrinter *AP, uint64_t LocCookie,
                                 raw_ostream &OS) {
-  int CurVariant = -1;            // The number of the {.|.|.} region we are in.
+  int CurVariant = -1; // The number of the {.|.|.} region we are in.
   const char *LastEmitted = AsmStr; // One past the last character emitted.
   unsigned NumOperands = MI->getNumOperands();
-  int AsmPrinterVariant = MAI->getAssemblerDialect();
+  int AsmPrinterVariant = MMI->getTarget().unqualifiedInlineAsmVariant();
 
   if (MAI->getEmitGNUAsmStartIndentationMarker())
     OS << '\t';
@@ -291,7 +324,7 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
              *LiteralEnd != '}' && *LiteralEnd != '$' && *LiteralEnd != '\n')
         ++LiteralEnd;
       if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
-        OS.write(LastEmitted, LiteralEnd-LastEmitted);
+        OS.write(LastEmitted, LiteralEnd - LastEmitted);
       LastEmitted = LiteralEnd;
       break;
     }
@@ -311,24 +344,24 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
           OS << '$';
         ++LastEmitted;  // Consume second '$' character.
         break;
-      case '(':             // $( -> same as GCC's { character.
-        ++LastEmitted;      // Consume '(' character.
+      case '(':        // $( -> same as GCC's { character.
+        ++LastEmitted; // Consume '(' character.
         if (CurVariant != -1)
           report_fatal_error("Nested variants found in inline asm string: '" +
                              Twine(AsmStr) + "'");
-        CurVariant = 0;     // We're in the first variant now.
+        CurVariant = 0; // We're in the first variant now.
         break;
       case '|':
-        ++LastEmitted;  // consume '|' character.
+        ++LastEmitted; // Consume '|' character.
         if (CurVariant == -1)
-          OS << '|';       // this is gcc's behavior for | outside a variant
+          OS << '|'; // This is gcc's behavior for | outside a variant.
         else
-          ++CurVariant;   // We're in the next variant.
+          ++CurVariant; // We're in the next variant.
         break;
-      case ')':         // $) -> same as GCC's } char.
-        ++LastEmitted;  // consume ')' character.
+      case ')':        // $) -> same as GCC's } char.
+        ++LastEmitted; // Consume ')' character.
         if (CurVariant == -1)
-          OS << '}';     // this is gcc's behavior for } outside a variant
+          OS << '}'; // This is gcc's behavior for } outside a variant.
         else
           CurVariant = -1;
         break;
@@ -351,9 +384,8 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
         if (!StrEnd)
           report_fatal_error("Unterminated ${:foo} operand in inline asm"
                              " string: '" + Twine(AsmStr) + "'");
-
-        std::string Val(StrStart, StrEnd);
-        AP->PrintSpecial(MI, OS, Val.c_str());
+        if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
+          AP->PrintSpecial(MI, OS, StringRef(StrStart, StrEnd - StrStart));
         LastEmitted = StrEnd+1;
         break;
       }
@@ -369,6 +401,10 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
                            Twine(AsmStr) + "'");
       LastEmitted = IDEnd;
 
+      if (Val >= NumOperands - 1)
+        report_fatal_error("Invalid $ operand number in inline asm string: '" +
+                           Twine(AsmStr) + "'");
+
       char Modifier[2] = { 0, 0 };
 
       if (HasCurlyBraces) {
@@ -390,10 +426,6 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
         ++LastEmitted;    // Consume '}' character.
       }
 
-      if (Val >= NumOperands-1)
-        report_fatal_error("Invalid $ operand number in inline asm string: '" +
-                           Twine(AsmStr) + "'");
-
       // Okay, we finally have a value number.  Ask the target to print this
       // operand!
       if (CurVariant == -1 || CurVariant == AsmPrinterVariant) {
@@ -403,7 +435,8 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
 
         // Scan to find the machine operand number for the operand.
         for (; Val; --Val) {
-          if (OpNo >= MI->getNumOperands()) break;
+          if (OpNo >= MI->getNumOperands())
+            break;
           unsigned OpFlags = MI->getOperand(OpNo).getImm();
           OpNo += InlineAsm::getNumOperandRegisters(OpFlags) + 1;
         }
@@ -411,12 +444,11 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
         // We may have a location metadata attached to the end of the
         // instruction, and at no point should see metadata at any
         // other point while processing. It's an error if so.
-        if (OpNo >= MI->getNumOperands() ||
-            MI->getOperand(OpNo).isMetadata()) {
+        if (OpNo >= MI->getNumOperands() || MI->getOperand(OpNo).isMetadata()) {
           Error = true;
         } else {
           unsigned OpFlags = MI->getOperand(OpNo).getImm();
-          ++OpNo;  // Skip over the ID number.
+          ++OpNo; // Skip over the ID number.
 
           // FIXME: Shouldn't arch-independent output template handling go into
           // PrintAsmOperand?
@@ -429,8 +461,6 @@ static void EmitGCCInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
           } else if (MI->getOperand(OpNo).isMBB()) {
             const MCSymbol *Sym = MI->getOperand(OpNo).getMBB()->getSymbol();
             Sym->print(OS, AP->MAI);
-          } else if (Modifier[0] == 'l') {
-            Error = true;
           } else if (InlineAsm::isMemKind(OpFlags)) {
             Error = AP->PrintAsmMemoryOperand(
                 MI, OpNo, Modifier[0] ? Modifier : nullptr, OS);
@@ -506,7 +536,7 @@ void AsmPrinter::emitInlineAsm(const MachineInstr *MI) const {
   if (MI->getInlineAsmDialect() == InlineAsm::AD_ATT)
     EmitGCCInlineAsmStr(AsmStr, MI, MMI, MAI, AP, LocCookie, OS);
   else
-    EmitMSInlineAsmStr(AsmStr, MI, MMI, AP, LocCookie, OS);
+    EmitMSInlineAsmStr(AsmStr, MI, MMI, MAI, AP, LocCookie, OS);
 
   // Emit warnings if we use reserved registers on the clobber list, as
   // that might lead to undefined behaviour.
@@ -540,7 +570,7 @@ void AsmPrinter::emitInlineAsm(const MachineInstr *MI) const {
         "preserved across the asm statement, and clobbering them may "
         "lead to undefined behaviour.";
     MMI->getModule()->getContext().diagnose(DiagnosticInfoInlineAsm(
-        LocCookie, Msg.c_str(), DiagnosticSeverity::DS_Warning));
+        LocCookie, Msg, DiagnosticSeverity::DS_Warning));
     MMI->getModule()->getContext().diagnose(
         DiagnosticInfoInlineAsm(LocCookie, Note, DiagnosticSeverity::DS_Note));
   }
@@ -560,13 +590,13 @@ void AsmPrinter::emitInlineAsm(const MachineInstr *MI) const {
 /// syntax used is ${:comment}.  Targets can override this to add support
 /// for their own strange codes.
 void AsmPrinter::PrintSpecial(const MachineInstr *MI, raw_ostream &OS,
-                              const char *Code) const {
-  if (!strcmp(Code, "private")) {
+                              StringRef Code) const {
+  if (Code == "private") {
     const DataLayout &DL = MF->getDataLayout();
     OS << DL.getPrivateGlobalPrefix();
-  } else if (!strcmp(Code, "comment")) {
+  } else if (Code == "comment") {
     OS << MAI->getCommentString();
-  } else if (!strcmp(Code, "uid")) {
+  } else if (Code == "uid") {
     // Comparing the address of MI isn't sufficient, because machineinstrs may
     // be allocated to the same address across functions.
 
@@ -582,7 +612,7 @@ void AsmPrinter::PrintSpecial(const MachineInstr *MI, raw_ostream &OS,
     raw_string_ostream Msg(msg);
     Msg << "Unknown special formatter '" << Code
          << "' for machine instr: " << *MI;
-    report_fatal_error(Msg.str());
+    report_fatal_error(Twine(Msg.str()));
   }
 }
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp b/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp
index bbb0504550c3..85ff84484ced 100644
--- a/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp
@@ -341,7 +341,16 @@ std::string CodeViewDebug::getFullyQualifiedName(const DIScope *Ty) {
 
 TypeIndex CodeViewDebug::getScopeIndex(const DIScope *Scope) {
   // No scope means global scope and that uses the zero index.
-  if (!Scope || isa<DIFile>(Scope))
+  //
+  // We also use zero index when the scope is a DISubprogram
+  // to suppress the emission of LF_STRING_ID for the function,
+  // which can trigger a link-time error with the linker in
+  // VS2019 version 16.11.2 or newer.
+  // Note, however, skipping the debug info emission for the DISubprogram
+  // is a temporary fix. The root issue here is that we need to figure out
+  // the proper way to encode a function nested in another function
+  // (as introduced by the Fortran 'contains' keyword) in CodeView.
+  if (!Scope || isa<DIFile>(Scope) || isa<DISubprogram>(Scope))
     return TypeIndex();
 
   assert(!isa<DIType>(Scope) && "shouldn't make a namespace scope for a type");
@@ -561,6 +570,44 @@ void CodeViewDebug::emitCodeViewMagicVersion() {
   OS.emitInt32(COFF::DEBUG_SECTION_MAGIC);
 }
 
+static SourceLanguage MapDWLangToCVLang(unsigned DWLang) {
+  switch (DWLang) {
+  case dwarf::DW_LANG_C:
+  case dwarf::DW_LANG_C89:
+  case dwarf::DW_LANG_C99:
+  case dwarf::DW_LANG_C11:
+  case dwarf::DW_LANG_ObjC:
+    return SourceLanguage::C;
+  case dwarf::DW_LANG_C_plus_plus:
+  case dwarf::DW_LANG_C_plus_plus_03:
+  case dwarf::DW_LANG_C_plus_plus_11:
+  case dwarf::DW_LANG_C_plus_plus_14:
+    return SourceLanguage::Cpp;
+  case dwarf::DW_LANG_Fortran77:
+  case dwarf::DW_LANG_Fortran90:
+  case dwarf::DW_LANG_Fortran95:
+  case dwarf::DW_LANG_Fortran03:
+  case dwarf::DW_LANG_Fortran08:
+    return SourceLanguage::Fortran;
+  case dwarf::DW_LANG_Pascal83:
+    return SourceLanguage::Pascal;
+  case dwarf::DW_LANG_Cobol74:
+  case dwarf::DW_LANG_Cobol85:
+    return SourceLanguage::Cobol;
+  case dwarf::DW_LANG_Java:
+    return SourceLanguage::Java;
+  case dwarf::DW_LANG_D:
+    return SourceLanguage::D;
+  case dwarf::DW_LANG_Swift:
+    return SourceLanguage::Swift;
+  default:
+    // There's no CodeView representation for this language, and CV doesn't
+    // have an "unknown" option for the language field, so we'll use MASM,
+    // as it's very low level.
+    return SourceLanguage::Masm;
+  }
+}
+
 void CodeViewDebug::beginModule(Module *M) {
   // If module doesn't have named metadata anchors or COFF debug section
   // is not available, skip any debug info related stuff.
@@ -574,6 +621,13 @@ void CodeViewDebug::beginModule(Module *M) {
 
   TheCPU = mapArchToCVCPUType(Triple(M->getTargetTriple()).getArch());
 
+  // Get the current source language.
+  NamedMDNode *CUs = MMI->getModule()->getNamedMetadata("llvm.dbg.cu");
+  const MDNode *Node = *CUs->operands().begin();
+  const auto *CU = cast<DICompileUnit>(Node);
+
+  CurrentSourceLanguage = MapDWLangToCVLang(CU->getSourceLanguage());
+
   collectGlobalVariableInfo();
 
   // Check if we should emit type record hashes.
@@ -731,43 +785,6 @@ void CodeViewDebug::emitTypeGlobalHashes() {
   }
 }
 
-static SourceLanguage MapDWLangToCVLang(unsigned DWLang) {
-  switch (DWLang) {
-  case dwarf::DW_LANG_C:
-  case dwarf::DW_LANG_C89:
-  case dwarf::DW_LANG_C99:
-  case dwarf::DW_LANG_C11:
-  case dwarf::DW_LANG_ObjC:
-    return SourceLanguage::C;
-  case dwarf::DW_LANG_C_plus_plus:
-  case dwarf::DW_LANG_C_plus_plus_03:
-  case dwarf::DW_LANG_C_plus_plus_11:
-  case dwarf::DW_LANG_C_plus_plus_14:
-    return SourceLanguage::Cpp;
-  case dwarf::DW_LANG_Fortran77:
-  case dwarf::DW_LANG_Fortran90:
-  case dwarf::DW_LANG_Fortran03:
-  case dwarf::DW_LANG_Fortran08:
-    return SourceLanguage::Fortran;
-  case dwarf::DW_LANG_Pascal83:
-    return SourceLanguage::Pascal;
-  case dwarf::DW_LANG_Cobol74:
-  case dwarf::DW_LANG_Cobol85:
-    return SourceLanguage::Cobol;
-  case dwarf::DW_LANG_Java:
-    return SourceLanguage::Java;
-  case dwarf::DW_LANG_D:
-    return SourceLanguage::D;
-  case dwarf::DW_LANG_Swift:
-    return SourceLanguage::Swift;
-  default:
-    // There's no CodeView representation for this language, and CV doesn't
-    // have an "unknown" option for the language field, so we'll use MASM,
-    // as it's very low level.
-    return SourceLanguage::Masm;
-  }
-}
-
 namespace {
 struct Version {
   int Part[4];
@@ -797,12 +814,8 @@ void CodeViewDebug::emitCompilerInformation() {
   MCSymbol *CompilerEnd = beginSymbolRecord(SymbolKind::S_COMPILE3);
   uint32_t Flags = 0;
 
-  NamedMDNode *CUs = MMI->getModule()->getNamedMetadata("llvm.dbg.cu");
-  const MDNode *Node = *CUs->operands().begin();
-  const auto *CU = cast<DICompileUnit>(Node);
-
   // The low byte of the flags indicates the source language.
-  Flags = MapDWLangToCVLang(CU->getSourceLanguage());
+  Flags = CurrentSourceLanguage;
   // TODO:  Figure out which other flags need to be set.
   if (MMI->getModule()->getProfileSummary(/*IsCS*/ false) != nullptr) {
     Flags |= static_cast<uint32_t>(CompileSym3Flags::PGO);
@@ -814,6 +827,10 @@ void CodeViewDebug::emitCompilerInformation() {
   OS.AddComment("CPUType");
   OS.emitInt16(static_cast<uint64_t>(TheCPU));
 
+  NamedMDNode *CUs = MMI->getModule()->getNamedMetadata("llvm.dbg.cu");
+  const MDNode *Node = *CUs->operands().begin();
+  const auto *CU = cast<DICompileUnit>(Node);
+
   StringRef CompilerVersion = CU->getProducer();
   Version FrontVer = parseVersion(CompilerVersion);
   OS.AddComment("Frontend version");
@@ -1573,6 +1590,8 @@ TypeIndex CodeViewDebug::lowerType(const DIType *Ty, const DIType *ClassTy) {
     return lowerTypeClass(cast<DICompositeType>(Ty));
   case dwarf::DW_TAG_union_type:
     return lowerTypeUnion(cast<DICompositeType>(Ty));
+  case dwarf::DW_TAG_string_type:
+    return lowerTypeString(cast<DIStringType>(Ty));
   case dwarf::DW_TAG_unspecified_type:
     if (Ty->getName() == "decltype(nullptr)")
       return TypeIndex::NullptrT();
@@ -1617,14 +1636,19 @@ TypeIndex CodeViewDebug::lowerTypeArray(const DICompositeType *Ty) {
 
     const DISubrange *Subrange = cast<DISubrange>(Element);
     int64_t Count = -1;
-    // Calculate the count if either LowerBound is absent or is zero and
-    // either of Count or UpperBound are constant.
-    auto *LI = Subrange->getLowerBound().dyn_cast<ConstantInt *>();
-    if (!Subrange->getRawLowerBound() || (LI && (LI->getSExtValue() == 0))) {
-      if (auto *CI = Subrange->getCount().dyn_cast<ConstantInt*>())
-        Count = CI->getSExtValue();
-      else if (auto *UI = Subrange->getUpperBound().dyn_cast<ConstantInt*>())
-        Count = UI->getSExtValue() + 1; // LowerBound is zero
+
+    // If Subrange has a Count field, use it.
+    // Otherwise, if it has an upperboud, use (upperbound - lowerbound + 1),
+    // where lowerbound is from the LowerBound field of the Subrange,
+    // or the language default lowerbound if that field is unspecified.
+    if (auto *CI = Subrange->getCount().dyn_cast<ConstantInt *>())
+      Count = CI->getSExtValue();
+    else if (auto *UI = Subrange->getUpperBound().dyn_cast<ConstantInt *>()) {
+      // Fortran uses 1 as the default lowerbound; other languages use 0.
+      int64_t Lowerbound = (moduleIsInFortran()) ? 1 : 0;
+      auto *LI = Subrange->getLowerBound().dyn_cast<ConstantInt *>();
+      Lowerbound = (LI) ? LI->getSExtValue() : Lowerbound;
+      Count = UI->getSExtValue() - Lowerbound + 1;
     }
 
     // Forward declarations of arrays without a size and VLAs use a count of -1.
@@ -1650,6 +1674,26 @@ TypeIndex CodeViewDebug::lowerTypeArray(const DICompositeType *Ty) {
   return ElementTypeIndex;
 }
 
+// This function lowers a Fortran character type (DIStringType).
+// Note that it handles only the character*n variant (using SizeInBits
+// field in DIString to describe the type size) at the moment.
+// Other variants (leveraging the StringLength and StringLengthExp
+// fields in DIStringType) remain TBD.
+TypeIndex CodeViewDebug::lowerTypeString(const DIStringType *Ty) {
+  TypeIndex CharType = TypeIndex(SimpleTypeKind::NarrowCharacter);
+  uint64_t ArraySize = Ty->getSizeInBits() >> 3;
+  StringRef Name = Ty->getName();
+  // IndexType is size_t, which depends on the bitness of the target.
+  TypeIndex IndexType = getPointerSizeInBytes() == 8
+                            ? TypeIndex(SimpleTypeKind::UInt64Quad)
+                            : TypeIndex(SimpleTypeKind::UInt32Long);
+
+  // Create a type of character array of ArraySize.
+  ArrayRecord AR(CharType, IndexType, ArraySize, Name);
+
+  return TypeTable.writeLeafType(AR);
+}
+
 TypeIndex CodeViewDebug::lowerTypeBasic(const DIBasicType *Ty) {
   TypeIndex Index;
   dwarf::TypeKind Kind;
@@ -1728,9 +1772,14 @@ TypeIndex CodeViewDebug::lowerTypeBasic(const DIBasicType *Ty) {
   }
 
   // Apply some fixups based on the source-level type name.
-  if (STK == SimpleTypeKind::Int32 && Ty->getName() == "long int")
+  // Include some amount of canonicalization from an old naming scheme Clang
+  // used to use for integer types (in an outdated effort to be compatible with
+  // GCC's debug info/GDB's behavior, which has since been addressed).
+  if (STK == SimpleTypeKind::Int32 &&
+      (Ty->getName() == "long int" || Ty->getName() == "long"))
     STK = SimpleTypeKind::Int32Long;
-  if (STK == SimpleTypeKind::UInt32 && Ty->getName() == "long unsigned int")
+  if (STK == SimpleTypeKind::UInt32 && (Ty->getName() == "long unsigned int" ||
+                                        Ty->getName() == "unsigned long"))
     STK = SimpleTypeKind::UInt32Long;
   if (STK == SimpleTypeKind::UInt16Short &&
       (Ty->getName() == "wchar_t" || Ty->getName() == "__wchar_t"))
@@ -2177,6 +2226,7 @@ void CodeViewDebug::clear() {
   TypeIndices.clear();
   CompleteTypeIndices.clear();
   ScopeGlobals.clear();
+  CVGlobalVariableOffsets.clear();
 }
 
 void CodeViewDebug::collectMemberInfo(ClassInfo &Info,
@@ -3062,6 +3112,15 @@ void CodeViewDebug::collectGlobalVariableInfo() {
       const DIGlobalVariable *DIGV = GVE->getVariable();
       const DIExpression *DIE = GVE->getExpression();
 
+      if ((DIE->getNumElements() == 2) &&
+          (DIE->getElement(0) == dwarf::DW_OP_plus_uconst))
+        // Record the constant offset for the variable.
+        //
+        // A Fortran common block uses this idiom to encode the offset
+        // of a variable from the common block's starting address.
+        CVGlobalVariableOffsets.insert(
+            std::make_pair(DIGV, DIE->getElement(1)));
+
       // Emit constant global variables in a global symbol section.
       if (GlobalMap.count(GVE) == 0 && DIE->isConstant()) {
         CVGlobalVariable CVGV = {DIGV, DIE};
@@ -3226,7 +3285,11 @@ void CodeViewDebug::emitDebugInfoForGlobal(const CVGlobalVariable &CVGV) {
   if (const auto *MemberDecl = dyn_cast_or_null<DIDerivedType>(
           DIGV->getRawStaticDataMemberDeclaration()))
     Scope = MemberDecl->getScope();
-  std::string QualifiedName = getFullyQualifiedName(Scope, DIGV->getName());
+  // For Fortran, the scoping portion is elided in its name so that we can
+  // reference the variable in the command line of the VS debugger.
+  std::string QualifiedName =
+      (moduleIsInFortran()) ? std::string(DIGV->getName())
+                            : getFullyQualifiedName(Scope, DIGV->getName());
 
   if (const GlobalVariable *GV =
           CVGV.GVInfo.dyn_cast<const GlobalVariable *>()) {
@@ -3242,7 +3305,13 @@ void CodeViewDebug::emitDebugInfoForGlobal(const CVGlobalVariable &CVGV) {
     OS.AddComment("Type");
     OS.emitInt32(getCompleteTypeIndex(DIGV->getType()).getIndex());
     OS.AddComment("DataOffset");
-    OS.EmitCOFFSecRel32(GVSym, /*Offset=*/0);
+
+    uint64_t Offset = 0;
+    if (CVGlobalVariableOffsets.find(DIGV) != CVGlobalVariableOffsets.end())
+      // Use the offset seen while collecting info on globals.
+      Offset = CVGlobalVariableOffsets[DIGV];
+    OS.EmitCOFFSecRel32(GVSym, Offset);
+
     OS.AddComment("Segment");
     OS.EmitCOFFSectionIndex(GVSym);
     OS.AddComment("Name");
diff --git a/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h b/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h
index d133474ee5aa..6f88e15ee8fe 100644
--- a/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h
+++ b/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.h
@@ -186,6 +186,13 @@ class LLVM_LIBRARY_VISIBILITY CodeViewDebug : public DebugHandlerBase {
   };
   FunctionInfo *CurFn = nullptr;
 
+  codeview::SourceLanguage CurrentSourceLanguage =
+      codeview::SourceLanguage::Masm;
+
+  // This map records the constant offset in DIExpression of the
+  // DIGlobalVariableExpression referencing the DIGlobalVariable.
+  DenseMap<const DIGlobalVariable *, uint64_t> CVGlobalVariableOffsets;
+
   // Map used to seperate variables according to the lexical scope they belong
   // in.  This is populated by recordLocalVariable() before
   // collectLexicalBlocks() separates the variables between the FunctionInfo
@@ -400,6 +407,7 @@ class LLVM_LIBRARY_VISIBILITY CodeViewDebug : public DebugHandlerBase {
   codeview::TypeIndex lowerType(const DIType *Ty, const DIType *ClassTy);
   codeview::TypeIndex lowerTypeAlias(const DIDerivedType *Ty);
   codeview::TypeIndex lowerTypeArray(const DICompositeType *Ty);
+  codeview::TypeIndex lowerTypeString(const DIStringType *Ty);
   codeview::TypeIndex lowerTypeBasic(const DIBasicType *Ty);
   codeview::TypeIndex lowerTypePointer(
       const DIDerivedType *Ty,
@@ -464,6 +472,11 @@ protected:
   /// Gather post-function debug information.
   void endFunctionImpl(const MachineFunction *) override;
 
+  /// Check if the current module is in Fortran.
+  bool moduleIsInFortran() {
+    return CurrentSourceLanguage == codeview::SourceLanguage::Fortran;
+  }
+
 public:
   CodeViewDebug(AsmPrinter *AP);
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/DIEHash.cpp b/llvm/lib/CodeGen/AsmPrinter/DIEHash.cpp
index 802f0e880514..5f4ee747fcca 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DIEHash.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DIEHash.cpp
@@ -93,19 +93,15 @@ void DIEHash::addParentContext(const DIE &Parent) {
 
   // Reverse iterate over our list to go from the outermost construct to the
   // innermost.
-  for (SmallVectorImpl<const DIE *>::reverse_iterator I = Parents.rbegin(),
-                                                      E = Parents.rend();
-       I != E; ++I) {
-    const DIE &Die = **I;
-
+  for (const DIE *Die : llvm::reverse(Parents)) {
     // ... Append the letter "C" to the sequence...
     addULEB128('C');
 
     // ... Followed by the DWARF tag of the construct...
-    addULEB128(Die.getTag());
+    addULEB128(Die->getTag());
 
     // ... Then the name, taken from the DW_AT_name attribute.
-    StringRef Name = getDIEStringAttr(Die, dwarf::DW_AT_name);
+    StringRef Name = getDIEStringAttr(*Die, dwarf::DW_AT_name);
     LLVM_DEBUG(dbgs() << "... adding context: " << Name << "\n");
     if (!Name.empty())
       addString(Name);
diff --git a/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp b/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp
index bb24f1414ef1..dd795079ac1a 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DbgEntityHistoryCalculator.cpp
@@ -252,8 +252,8 @@ void DbgValueHistoryMap::trimLocationRanges(
 
     // Now actually remove the entries. Iterate backwards so that our remaining
     // ToRemove indices are valid after each erase.
-    for (auto Itr = ToRemove.rbegin(), End = ToRemove.rend(); Itr != End; ++Itr)
-      HistoryMapEntries.erase(HistoryMapEntries.begin() + *Itr);
+    for (EntryIndex Idx : llvm::reverse(ToRemove))
+      HistoryMapEntries.erase(HistoryMapEntries.begin() + Idx);
   }
 }
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/DebugHandlerBase.cpp b/llvm/lib/CodeGen/AsmPrinter/DebugHandlerBase.cpp
index c81288c0e460..4df34d2c9402 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DebugHandlerBase.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DebugHandlerBase.cpp
@@ -174,21 +174,26 @@ uint64_t DebugHandlerBase::getBaseTypeSize(const DIType *Ty) {
 }
 
 bool DebugHandlerBase::isUnsignedDIType(const DIType *Ty) {
-  // SROA may generate dbg value intrinsics to assign an unsigned value to a
-  // Fortran CHARACTER(1) type variables. Make them as unsigned.
   if (isa<DIStringType>(Ty)) {
-    assert((Ty->getSizeInBits()) == 8 && "Not a valid unsigned type!");
+    // Some transformations (e.g. instcombine) may decide to turn a Fortran
+    // character object into an integer, and later ones (e.g. SROA) may
+    // further inject a constant integer in a llvm.dbg.value call to track
+    // the object's value. Here we trust the transformations are doing the
+    // right thing, and treat the constant as unsigned to preserve that value
+    // (i.e. avoid sign extension).
     return true;
   }
-  if (auto *CTy = dyn_cast<DICompositeType>(Ty)) {
-    // FIXME: Enums without a fixed underlying type have unknown signedness
-    // here, leading to incorrectly emitted constants.
-    if (CTy->getTag() == dwarf::DW_TAG_enumeration_type)
-      return false;
 
-    // (Pieces of) aggregate types that get hacked apart by SROA may be
-    // represented by a constant. Encode them as unsigned bytes.
-    return true;
+  if (auto *CTy = dyn_cast<DICompositeType>(Ty)) {
+    if (CTy->getTag() == dwarf::DW_TAG_enumeration_type) {
+      if (!(Ty = CTy->getBaseType()))
+        // FIXME: Enums without a fixed underlying type have unknown signedness
+        // here, leading to incorrectly emitted constants.
+        return false;
+    } else
+      // (Pieces of) aggregate types that get hacked apart by SROA may be
+      // represented by a constant. Encode them as unsigned bytes.
+      return true;
   }
 
   if (auto *DTy = dyn_cast<DIDerivedType>(Ty)) {
diff --git a/llvm/lib/CodeGen/AsmPrinter/DebugLocEntry.h b/llvm/lib/CodeGen/AsmPrinter/DebugLocEntry.h
index 62ebadaf3cbe..d7ab2091967f 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DebugLocEntry.h
+++ b/llvm/lib/CodeGen/AsmPrinter/DebugLocEntry.h
@@ -158,7 +158,7 @@ public:
   friend bool operator<(const DbgValueLoc &, const DbgValueLoc &);
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
   LLVM_DUMP_METHOD void dump() const {
-    for (DbgValueLocEntry DV : ValueLocEntries)
+    for (const DbgValueLocEntry &DV : ValueLocEntries)
       DV.dump();
     if (Expression)
       Expression->dump();
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
index faa14dca1c3f..922c91840520 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
@@ -143,8 +143,6 @@ DIE *DwarfCompileUnit::getOrCreateGlobalVariableDIE(
   auto *GVContext = GV->getScope();
   const DIType *GTy = GV->getType();
 
-  // Construct the context before querying for the existence of the DIE in
-  // case such construction creates the DIE.
   auto *CB = GVContext ? dyn_cast<DICommonBlock>(GVContext) : nullptr;
   DIE *ContextDIE = CB ? getOrCreateCommonBlock(CB, GlobalExprs)
     : getOrCreateContextDIE(GVContext);
@@ -183,6 +181,8 @@ DIE *DwarfCompileUnit::getOrCreateGlobalVariableDIE(
   else
     addGlobalName(GV->getName(), *VariableDIE, DeclContext);
 
+  addAnnotation(*VariableDIE, GV->getAnnotations());
+
   if (uint32_t AlignInBytes = GV->getAlignInBytes())
     addUInt(*VariableDIE, dwarf::DW_AT_alignment, dwarf::DW_FORM_udata,
             AlignInBytes);
@@ -260,14 +260,14 @@ void DwarfCompileUnit::addLocationAttribute(
 
     if (Global) {
       const MCSymbol *Sym = Asm->getSymbol(Global);
+      unsigned PointerSize = Asm->getDataLayout().getPointerSize();
+      assert((PointerSize == 4 || PointerSize == 8) &&
+             "Add support for other sizes if necessary");
       if (Global->isThreadLocal()) {
         if (Asm->TM.useEmulatedTLS()) {
           // TODO: add debug info for emulated thread local mode.
         } else {
           // FIXME: Make this work with -gsplit-dwarf.
-          unsigned PointerSize = Asm->getDataLayout().getPointerSize();
-          assert((PointerSize == 4 || PointerSize == 8) &&
-                 "Add support for other sizes if necessary");
           // Based on GCC's support for TLS:
           if (!DD->useSplitDwarf()) {
             // 1) Start with a constNu of the appropriate pointer size
@@ -290,6 +290,24 @@ void DwarfCompileUnit::addLocationAttribute(
                   DD->useGNUTLSOpcode() ? dwarf::DW_OP_GNU_push_tls_address
                                         : dwarf::DW_OP_form_tls_address);
         }
+      } else if (Asm->TM.getRelocationModel() == Reloc::RWPI ||
+                 Asm->TM.getRelocationModel() == Reloc::ROPI_RWPI) {
+        // Constant
+        addUInt(*Loc, dwarf::DW_FORM_data1,
+                PointerSize == 4 ? dwarf::DW_OP_const4u
+                                 : dwarf::DW_OP_const8u);
+        // Relocation offset
+        addExpr(*Loc, PointerSize == 4 ? dwarf::DW_FORM_data4
+                                       : dwarf::DW_FORM_data8,
+                Asm->getObjFileLowering().getIndirectSymViaRWPI(Sym));
+        // Base register
+        Register BaseReg = Asm->getObjFileLowering().getStaticBase();
+        BaseReg = Asm->TM.getMCRegisterInfo()->getDwarfRegNum(BaseReg, false);
+        addUInt(*Loc, dwarf::DW_FORM_data1, dwarf::DW_OP_breg0 + BaseReg);
+        // Offset from base register
+        addSInt(*Loc, dwarf::DW_FORM_sdata, 0);
+        // Operation
+        addUInt(*Loc, dwarf::DW_FORM_data1, dwarf::DW_OP_plus);
       } else {
         DD->addArangeLabel(SymbolCU(this, Sym));
         addOpAddress(*Loc, Sym);
@@ -331,12 +349,10 @@ void DwarfCompileUnit::addLocationAttribute(
 
 DIE *DwarfCompileUnit::getOrCreateCommonBlock(
     const DICommonBlock *CB, ArrayRef<GlobalExpr> GlobalExprs) {
-  // Construct the context before querying for the existence of the DIE in case
-  // such construction creates the DIE.
-  DIE *ContextDIE = getOrCreateContextDIE(CB->getScope());
-
+  // Check for pre-existence.
   if (DIE *NDie = getDIE(CB))
     return NDie;
+  DIE *ContextDIE = getOrCreateContextDIE(CB->getScope());
   DIE &NDie = createAndAddDIE(dwarf::DW_TAG_common_block, *ContextDIE, CB);
   StringRef Name = CB->getName().empty() ? "_BLNK_" : CB->getName();
   addString(NDie, dwarf::DW_AT_name, Name);
@@ -351,7 +367,8 @@ DIE *DwarfCompileUnit::getOrCreateCommonBlock(
 void DwarfCompileUnit::addRange(RangeSpan Range) {
   DD->insertSectionLabel(Range.Begin);
 
-  bool SameAsPrevCU = this == DD->getPrevCU();
+  auto *PrevCU = DD->getPrevCU();
+  bool SameAsPrevCU = this == PrevCU;
   DD->setPrevCU(this);
   // If we have no current ranges just add the range and return, otherwise,
   // check the current section and CU against the previous section and CU we
@@ -360,6 +377,9 @@ void DwarfCompileUnit::addRange(RangeSpan Range) {
   if (CURanges.empty() || !SameAsPrevCU ||
       (&CURanges.back().End->getSection() !=
        &Range.End->getSection())) {
+    // Before a new range is added, always terminate the prior line table.
+    if (PrevCU)
+      DD->terminateLineTable(PrevCU);
     CURanges.push_back(Range);
     return;
   }
@@ -470,7 +490,6 @@ DIE &DwarfCompileUnit::updateSubprogramScopeDIE(const DISubprogram *SP) {
         addSInt(*Loc, dwarf::DW_FORM_sdata, TI_GLOBAL_RELOC);
         if (!isDwoUnit()) {
           addLabel(*Loc, dwarf::DW_FORM_data4, SPSym);
-          DD->addArangeLabel(SymbolCU(this, SPSym));
         } else {
           // FIXME: when writing dwo, we need to avoid relocations. Probably
           // the "right" solution is to treat globals the way func and data
@@ -961,9 +980,7 @@ sortLocalVars(SmallVectorImpl<DbgVariable *> &Input) {
     bool visitedAllDependencies = Item.getInt();
     WorkList.pop_back();
 
-    // Dependency is in a different lexical scope or a global.
-    if (!Var)
-      continue;
+    assert(Var);
 
     // Already handled.
     if (Visited.count(Var))
@@ -987,8 +1004,10 @@ sortLocalVars(SmallVectorImpl<DbgVariable *> &Input) {
     // visited again after all of its dependencies are handled.
     WorkList.push_back({Var, 1});
     for (auto *Dependency : dependencies(Var)) {
-      auto Dep = dyn_cast_or_null<const DILocalVariable>(Dependency);
-      WorkList.push_back({DbgVar[Dep], 0});
+      // Don't add dependency if it is in a different lexical scope or a global.
+      if (const auto *Dep = dyn_cast<const DILocalVariable>(Dependency))
+        if (DbgVariable *Var = DbgVar.lookup(Dep))
+          WorkList.push_back({Var, 0});
     }
   }
   return Result;
@@ -1103,9 +1122,10 @@ void DwarfCompileUnit::constructAbstractSubprogramScopeDIE(
   // shouldn't be found by lookup.
   AbsDef = &ContextCU->createAndAddDIE(dwarf::DW_TAG_subprogram, *ContextDIE, nullptr);
   ContextCU->applySubprogramAttributesToDefinition(SP, *AbsDef);
-
-  if (!ContextCU->includeMinimalInlineScopes())
-    ContextCU->addUInt(*AbsDef, dwarf::DW_AT_inline, None, dwarf::DW_INL_inlined);
+  ContextCU->addSInt(*AbsDef, dwarf::DW_AT_inline,
+                     DD->getDwarfVersion() <= 4 ? Optional<dwarf::Form>()
+                                                : dwarf::DW_FORM_implicit_const,
+                     dwarf::DW_INL_inlined);
   if (DIE *ObjectPointer = ContextCU->createAndAddScopeChildren(Scope, *AbsDef))
     ContextCU->addDIEEntry(*AbsDef, dwarf::DW_AT_object_pointer, *ObjectPointer);
 }
@@ -1162,7 +1182,7 @@ DwarfCompileUnit::getDwarf5OrGNULocationAtom(dwarf::LocationAtom Loc) const {
 }
 
 DIE &DwarfCompileUnit::constructCallSiteEntryDIE(DIE &ScopeDIE,
-                                                 DIE *CalleeDIE,
+                                                 const DISubprogram *CalleeSP,
                                                  bool IsTail,
                                                  const MCSymbol *PCAddr,
                                                  const MCSymbol *CallAddr,
@@ -1176,7 +1196,8 @@ DIE &DwarfCompileUnit::constructCallSiteEntryDIE(DIE &ScopeDIE,
     addAddress(CallSiteDIE, getDwarf5OrGNUAttr(dwarf::DW_AT_call_target),
                MachineLocation(CallReg));
   } else {
-    assert(CalleeDIE && "No DIE for call site entry origin");
+    DIE *CalleeDIE = getOrCreateSubprogramDIE(CalleeSP);
+    assert(CalleeDIE && "Could not create DIE for call site entry origin");
     addDIEEntry(CallSiteDIE, getDwarf5OrGNUAttr(dwarf::DW_AT_call_origin),
                 *CalleeDIE);
   }
@@ -1265,6 +1286,16 @@ DIE *DwarfCompileUnit::constructImportedEntityDIE(
   if (!Name.empty())
     addString(*IMDie, dwarf::DW_AT_name, Name);
 
+  // This is for imported module with renamed entities (such as variables and
+  // subprograms).
+  DINodeArray Elements = Module->getElements();
+  for (const auto *Element : Elements) {
+    if (!Element)
+      continue;
+    IMDie->addChild(
+        constructImportedEntityDIE(cast<DIImportedEntity>(Element)));
+  }
+
   return IMDie;
 }
 
@@ -1479,10 +1510,12 @@ void DwarfCompileUnit::applyVariableAttributes(const DbgVariable &Var,
   if (!Name.empty())
     addString(VariableDie, dwarf::DW_AT_name, Name);
   const auto *DIVar = Var.getVariable();
-  if (DIVar)
+  if (DIVar) {
     if (uint32_t AlignInBytes = DIVar->getAlignInBytes())
       addUInt(VariableDie, dwarf::DW_AT_alignment, dwarf::DW_FORM_udata,
               AlignInBytes);
+    addAnnotation(VariableDie, DIVar->getAnnotations());
+  }
 
   addSourceLine(VariableDie, DIVar);
   addType(VariableDie, Var.getType());
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
index 6d8186a5ee2b..6e9261087686 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfCompileUnit.h
@@ -249,16 +249,14 @@ public:
   dwarf::LocationAtom getDwarf5OrGNULocationAtom(dwarf::LocationAtom Loc) const;
 
   /// Construct a call site entry DIE describing a call within \p Scope to a
-  /// callee described by \p CalleeDIE.
-  /// \p CalleeDIE is a declaration or definition subprogram DIE for the callee.
-  /// For indirect calls \p CalleeDIE is set to nullptr.
+  /// callee described by \p CalleeSP.
   /// \p IsTail specifies whether the call is a tail call.
   /// \p PCAddr points to the PC value after the call instruction.
   /// \p CallAddr points to the PC value at the call instruction (or is null).
   /// \p CallReg is a register location for an indirect call. For direct calls
   /// the \p CallReg is set to 0.
-  DIE &constructCallSiteEntryDIE(DIE &ScopeDIE, DIE *CalleeDIE, bool IsTail,
-                                 const MCSymbol *PCAddr,
+  DIE &constructCallSiteEntryDIE(DIE &ScopeDIE, const DISubprogram *CalleeSP,
+                                 bool IsTail, const MCSymbol *PCAddr,
                                  const MCSymbol *CallAddr, unsigned CallReg);
   /// Construct call site parameter DIEs for the \p CallSiteDIE. The \p Params
   /// were collected by the \ref collectCallSiteParameters.
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
index ee14423ca3d0..047676d4c11e 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
@@ -480,7 +480,7 @@ static bool hasObjCCategory(StringRef Name) {
   if (!isObjCClass(Name))
     return false;
 
-  return Name.find(") ") != StringRef::npos;
+  return Name.contains(") ");
 }
 
 static void getObjCClassCategory(StringRef In, StringRef &Class,
@@ -587,14 +587,6 @@ void DwarfDebug::constructAbstractSubprogramScopeDIE(DwarfCompileUnit &SrcCU,
   }
 }
 
-DIE &DwarfDebug::constructSubprogramDefinitionDIE(const DISubprogram *SP) {
-  DICompileUnit *Unit = SP->getUnit();
-  assert(SP->isDefinition() && "Subprogram not a definition");
-  assert(Unit && "Subprogram definition without parent unit");
-  auto &CU = getOrCreateDwarfCompileUnit(Unit);
-  return *CU.getOrCreateSubprogramDIE(SP);
-}
-
 /// Represents a parameter whose call site value can be described by applying a
 /// debug expression to a register in the forwarded register worklist.
 struct FwdRegParamInfo {
@@ -945,7 +937,7 @@ void DwarfDebug::constructCallSiteEntryDIEs(const DISubprogram &SP,
         continue;
 
       unsigned CallReg = 0;
-      DIE *CalleeDIE = nullptr;
+      const DISubprogram *CalleeSP = nullptr;
       const Function *CalleeDecl = nullptr;
       if (CalleeOp.isReg()) {
         CallReg = CalleeOp.getReg();
@@ -955,19 +947,7 @@ void DwarfDebug::constructCallSiteEntryDIEs(const DISubprogram &SP,
         CalleeDecl = dyn_cast<Function>(CalleeOp.getGlobal());
         if (!CalleeDecl || !CalleeDecl->getSubprogram())
           continue;
-        const DISubprogram *CalleeSP = CalleeDecl->getSubprogram();
-
-        if (CalleeSP->isDefinition()) {
-          // Ensure that a subprogram DIE for the callee is available in the
-          // appropriate CU.
-          CalleeDIE = &constructSubprogramDefinitionDIE(CalleeSP);
-        } else {
-          // Create the declaration DIE if it is missing. This is required to
-          // support compilation of old bitcode with an incomplete list of
-          // retained metadata.
-          CalleeDIE = CU.getOrCreateSubprogramDIE(CalleeSP);
-        }
-        assert(CalleeDIE && "Must have a DIE for the callee");
+        CalleeSP = CalleeDecl->getSubprogram();
       }
 
       // TODO: Omit call site entries for runtime calls (objc_msgSend, etc).
@@ -1004,7 +984,7 @@ void DwarfDebug::constructCallSiteEntryDIEs(const DISubprogram &SP,
                         << (IsTail ? " [IsTail]" : "") << "\n");
 
       DIE &CallSiteDIE = CU.constructCallSiteEntryDIE(
-          ScopeDIE, CalleeDIE, IsTail, PCAddr, CallAddr, CallReg);
+          ScopeDIE, CalleeSP, IsTail, PCAddr, CallAddr, CallReg);
 
       // Optionally emit call-site-param debug info.
       if (emitDebugEntryValues()) {
@@ -1427,6 +1407,10 @@ void DwarfDebug::finalizeModuleInfo() {
 
 // Emit all Dwarf sections that should come after the content.
 void DwarfDebug::endModule() {
+  // Terminate the pending line table.
+  if (PrevCU)
+    terminateLineTable(PrevCU);
+  PrevCU = nullptr;
   assert(CurFn == nullptr);
   assert(CurMI == nullptr);
 
@@ -2102,12 +2086,22 @@ void DwarfDebug::beginInstruction(const MachineInstr *MI) {
 static DebugLoc findPrologueEndLoc(const MachineFunction *MF) {
   // First known non-DBG_VALUE and non-frame setup location marks
   // the beginning of the function body.
-  for (const auto &MBB : *MF)
-    for (const auto &MI : MBB)
+  DebugLoc LineZeroLoc;
+  for (const auto &MBB : *MF) {
+    for (const auto &MI : MBB) {
       if (!MI.isMetaInstruction() && !MI.getFlag(MachineInstr::FrameSetup) &&
-          MI.getDebugLoc())
-        return MI.getDebugLoc();
-  return DebugLoc();
+          MI.getDebugLoc()) {
+        // Scan forward to try to find a non-zero line number. The prologue_end
+        // marks the first breakpoint in the function after the frame setup, and
+        // a compiler-generated line 0 location is not a meaningful breakpoint.
+        // If none is found, return the first location after the frame setup.
+        if (MI.getDebugLoc().getLine())
+          return MI.getDebugLoc();
+        LineZeroLoc = MI.getDebugLoc();
+      }
+    }
+  }
+  return LineZeroLoc;
 }
 
 /// Register a source line with debug info. Returns the  unique label that was
@@ -2162,24 +2156,42 @@ void DwarfDebug::beginFunctionImpl(const MachineFunction *MF) {
 
   DwarfCompileUnit &CU = getOrCreateDwarfCompileUnit(SP->getUnit());
 
+  Asm->OutStreamer->getContext().setDwarfCompileUnitID(
+      getDwarfCompileUnitIDForLineTable(CU));
+
+  // Record beginning of function.
+  PrologEndLoc = emitInitialLocDirective(
+      *MF, Asm->OutStreamer->getContext().getDwarfCompileUnitID());
+}
+
+unsigned
+DwarfDebug::getDwarfCompileUnitIDForLineTable(const DwarfCompileUnit &CU) {
   // Set DwarfDwarfCompileUnitID in MCContext to the Compile Unit this function
   // belongs to so that we add to the correct per-cu line table in the
   // non-asm case.
   if (Asm->OutStreamer->hasRawTextSupport())
     // Use a single line table if we are generating assembly.
-    Asm->OutStreamer->getContext().setDwarfCompileUnitID(0);
+    return 0;
   else
-    Asm->OutStreamer->getContext().setDwarfCompileUnitID(CU.getUniqueID());
+    return CU.getUniqueID();
+}
 
-  // Record beginning of function.
-  PrologEndLoc = emitInitialLocDirective(
-      *MF, Asm->OutStreamer->getContext().getDwarfCompileUnitID());
+void DwarfDebug::terminateLineTable(const DwarfCompileUnit *CU) {
+  const auto &CURanges = CU->getRanges();
+  auto &LineTable = Asm->OutStreamer->getContext().getMCDwarfLineTable(
+      getDwarfCompileUnitIDForLineTable(*CU));
+  // Add the last range label for the given CU.
+  LineTable.getMCLineSections().addEndEntry(
+      const_cast<MCSymbol *>(CURanges.back().End));
 }
 
 void DwarfDebug::skippedNonDebugFunction() {
   // If we don't have a subprogram for this function then there will be a hole
   // in the range information. Keep note of this by setting the previously used
   // section to nullptr.
+  // Terminate the pending line table.
+  if (PrevCU)
+    terminateLineTable(PrevCU);
   PrevCU = nullptr;
   CurFn = nullptr;
 }
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.h b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.h
index 6356a65b50d3..4e1a1b1e068d 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.h
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfDebug.h
@@ -65,19 +65,21 @@ class Module;
 /// such that it could levarage polymorphism to extract common code for
 /// DbgVariable and DbgLabel.
 class DbgEntity {
-  const DINode *Entity;
-  const DILocation *InlinedAt;
-  DIE *TheDIE = nullptr;
-  unsigned SubclassID;
-
 public:
   enum DbgEntityKind {
     DbgVariableKind,
     DbgLabelKind
   };
 
-  DbgEntity(const DINode *N, const DILocation *IA, unsigned ID)
-    : Entity(N), InlinedAt(IA), SubclassID(ID) {}
+private:
+  const DINode *Entity;
+  const DILocation *InlinedAt;
+  DIE *TheDIE = nullptr;
+  const DbgEntityKind SubclassID;
+
+public:
+  DbgEntity(const DINode *N, const DILocation *IA, DbgEntityKind ID)
+      : Entity(N), InlinedAt(IA), SubclassID(ID) {}
   virtual ~DbgEntity() {}
 
   /// Accessors.
@@ -85,19 +87,18 @@ public:
   const DINode *getEntity() const { return Entity; }
   const DILocation *getInlinedAt() const { return InlinedAt; }
   DIE *getDIE() const { return TheDIE; }
-  unsigned getDbgEntityID() const { return SubclassID; }
+  DbgEntityKind getDbgEntityID() const { return SubclassID; }
   /// @}
 
   void setDIE(DIE &D) { TheDIE = &D; }
 
   static bool classof(const DbgEntity *N) {
     switch (N->getDbgEntityID()) {
-    default:
-      return false;
     case DbgVariableKind:
     case DbgLabelKind:
       return true;
     }
+    llvm_unreachable("Invalid DbgEntityKind");
   }
 };
 
@@ -471,9 +472,6 @@ private:
   /// Construct a DIE for this abstract scope.
   void constructAbstractSubprogramScopeDIE(DwarfCompileUnit &SrcCU, LexicalScope *Scope);
 
-  /// Construct a DIE for the subprogram definition \p SP and return it.
-  DIE &constructSubprogramDefinitionDIE(const DISubprogram *SP);
-
   /// Construct DIEs for call site entries describing the calls in \p MF.
   void constructCallSiteEntryDIEs(const DISubprogram &SP, DwarfCompileUnit &CU,
                                   DIE &ScopeDIE, const MachineFunction &MF);
@@ -615,7 +613,7 @@ private:
                          DenseSet<InlinedEntity> &ProcessedVars);
 
   /// Build the location list for all DBG_VALUEs in the
-  /// function that describe the same variable. If the resulting 
+  /// function that describe the same variable. If the resulting
   /// list has only one entry that is valid for entire variable's
   /// scope return true.
   bool buildLocationList(SmallVectorImpl<DebugLocEntry> &DebugLoc,
@@ -635,6 +633,9 @@ protected:
   /// Gather and emit post-function debug information.
   void endFunctionImpl(const MachineFunction *MF) override;
 
+  /// Get Dwarf compile unit ID for line table.
+  unsigned getDwarfCompileUnitIDForLineTable(const DwarfCompileUnit &CU);
+
   void skippedNonDebugFunction() override;
 
 public:
@@ -781,6 +782,9 @@ public:
   const DwarfCompileUnit *getPrevCU() const { return PrevCU; }
   void setPrevCU(const DwarfCompileUnit *PrevCU) { this->PrevCU = PrevCU; }
 
+  /// Terminate the line table by adding the last range label.
+  void terminateLineTable(const DwarfCompileUnit *CU);
+
   /// Returns the entries for the .debug_loc section.
   const DebugLocStream &getDebugLocs() const { return DebugLocs; }
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp b/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp
index 344d30fad347..976e35905144 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp
@@ -186,9 +186,8 @@ int64_t DwarfUnit::getDefaultLowerBound() const {
 
 /// Check whether the DIE for this MDNode can be shared across CUs.
 bool DwarfUnit::isShareableAcrossCUs(const DINode *D) const {
-  // When the MDNode can be part of the type system (this includes subprogram
-  // declarations *and* subprogram definitions, even local definitions), the
-  // DIE must be shared across CUs.
+  // When the MDNode can be part of the type system, the DIE can be shared
+  // across CUs.
   // Combining type units and cross-CU DIE sharing is lower value (since
   // cross-CU DIE sharing is used in LTO and removes type redundancy at that
   // level already) but may be implementable for some value in projects
@@ -196,7 +195,9 @@ bool DwarfUnit::isShareableAcrossCUs(const DINode *D) const {
   // together.
   if (isDwoUnit() && !DD->shareAcrossDWOCUs())
     return false;
-  return (isa<DIType>(D) || isa<DISubprogram>(D)) && !DD->generateTypeUnits();
+  return (isa<DIType>(D) ||
+          (isa<DISubprogram>(D) && !cast<DISubprogram>(D)->isDefinition())) &&
+         !DD->generateTypeUnits();
 }
 
 DIE *DwarfUnit::getDIE(const DINode *D) const {
@@ -671,7 +672,7 @@ std::string DwarfUnit::getParentContextString(const DIScope *Context) const {
 
   // Reverse iterate over our list to go from the outermost construct to the
   // innermost.
-  for (const DIScope *Ctx : make_range(Parents.rbegin(), Parents.rend())) {
+  for (const DIScope *Ctx : llvm::reverse(Parents)) {
     StringRef Name = Ctx->getName();
     if (Name.empty() && isa<DINamespace>(Ctx))
       Name = "(anonymous namespace)";
@@ -753,6 +754,8 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DIDerivedType *DTy) {
   if (!Name.empty())
     addString(Buffer, dwarf::DW_AT_name, Name);
 
+  addAnnotation(Buffer, DTy->getAnnotations());
+
   // If alignment is specified for a typedef , create and insert DW_AT_alignment
   // attribute in DW_TAG_typedef DIE.
   if (Tag == dwarf::DW_TAG_typedef && DD->getDwarfVersion() >= 5) {
@@ -832,6 +835,23 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DISubroutineType *CTy) {
     addFlag(Buffer, dwarf::DW_AT_rvalue_reference);
 }
 
+void DwarfUnit::addAnnotation(DIE &Buffer, DINodeArray Annotations) {
+  if (!Annotations)
+    return;
+
+  for (const Metadata *Annotation : Annotations->operands()) {
+    const MDNode *MD = cast<MDNode>(Annotation);
+    const MDString *Name = cast<MDString>(MD->getOperand(0));
+
+    // Currently, only MDString is supported with btf_decl_tag attribute.
+    const MDString *Value = cast<MDString>(MD->getOperand(1));
+
+    DIE &AnnotationDie = createAndAddDIE(dwarf::DW_TAG_LLVM_annotation, Buffer);
+    addString(AnnotationDie, dwarf::DW_AT_name, Name->getString());
+    addString(AnnotationDie, dwarf::DW_AT_const_value, Value->getString());
+  }
+}
+
 void DwarfUnit::constructTypeDIE(DIE &Buffer, const DICompositeType *CTy) {
   // Add name if not anonymous or intermediate type.
   StringRef Name = CTy->getName();
@@ -849,7 +869,8 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DICompositeType *CTy) {
   case dwarf::DW_TAG_variant_part:
   case dwarf::DW_TAG_structure_type:
   case dwarf::DW_TAG_union_type:
-  case dwarf::DW_TAG_class_type: {
+  case dwarf::DW_TAG_class_type:
+  case dwarf::DW_TAG_namelist: {
     // Emit the discriminator for a variant part.
     DIDerivedType *Discriminator = nullptr;
     if (Tag == dwarf::DW_TAG_variant_part) {
@@ -918,6 +939,13 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DICompositeType *CTy) {
           DIE &VariantPart = createAndAddDIE(Composite->getTag(), Buffer);
           constructTypeDIE(VariantPart, Composite);
         }
+      } else if (Tag == dwarf::DW_TAG_namelist) {
+        auto *Var = dyn_cast<DINode>(Element);
+        auto *VarDIE = getDIE(Var);
+        if (VarDIE) {
+          DIE &ItemDie = createAndAddDIE(dwarf::DW_TAG_namelist_item, Buffer);
+          addDIEEntry(ItemDie, dwarf::DW_AT_namelist_item, *VarDIE);
+        }
       }
     }
 
@@ -960,6 +988,8 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DICompositeType *CTy) {
   if (!Name.empty())
     addString(Buffer, dwarf::DW_AT_name, Name);
 
+  addAnnotation(Buffer, CTy->getAnnotations());
+
   if (Tag == dwarf::DW_TAG_enumeration_type ||
       Tag == dwarf::DW_TAG_class_type || Tag == dwarf::DW_TAG_structure_type ||
       Tag == dwarf::DW_TAG_union_type) {
@@ -1196,6 +1226,8 @@ void DwarfUnit::applySubprogramAttributes(const DISubprogram *SP, DIE &SPDie,
   if (!SP->getName().empty())
     addString(SPDie, dwarf::DW_AT_name, SP->getName());
 
+  addAnnotation(SPDie, SP->getAnnotations());
+
   if (!SkipSPSourceLocation)
     addSourceLine(SPDie, SP);
 
@@ -1546,6 +1578,8 @@ DIE &DwarfUnit::constructMemberDIE(DIE &Buffer, const DIDerivedType *DT) {
   if (!Name.empty())
     addString(MemberDie, dwarf::DW_AT_name, Name);
 
+  addAnnotation(MemberDie, DT->getAnnotations());
+
   if (DIType *Resolved = DT->getBaseType())
     addType(MemberDie, Resolved);
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.h b/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.h
index 4d31dd0daf59..8140279adaef 100644
--- a/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.h
+++ b/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.h
@@ -294,6 +294,9 @@ public:
   void addSectionLabel(DIE &Die, dwarf::Attribute Attribute,
                        const MCSymbol *Label, const MCSymbol *Sec);
 
+  /// Add DW_TAG_LLVM_annotation.
+  void addAnnotation(DIE &Buffer, DINodeArray Annotations);
+
   /// Get context owner's DIE.
   DIE *createTypeDIE(const DICompositeType *Ty);
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/EHStreamer.cpp b/llvm/lib/CodeGen/AsmPrinter/EHStreamer.cpp
index e589c2e64abd..150f19324834 100644
--- a/llvm/lib/CodeGen/AsmPrinter/EHStreamer.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/EHStreamer.cpp
@@ -812,8 +812,7 @@ void EHStreamer::emitTypeInfos(unsigned TTypeEncoding, MCSymbol *TTBaseLabel) {
     Entry = TypeInfos.size();
   }
 
-  for (const GlobalValue *GV : make_range(TypeInfos.rbegin(),
-                                          TypeInfos.rend())) {
+  for (const GlobalValue *GV : llvm::reverse(TypeInfos)) {
     if (VerboseAsm)
       Asm->OutStreamer->AddComment("TypeInfo " + Twine(Entry--));
     Asm->emitTTypeReference(GV, TTypeEncoding);
diff --git a/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.cpp b/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.cpp
index 35a830f416f6..9e6f1a537de3 100644
--- a/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.cpp
@@ -20,6 +20,8 @@
 
 using namespace llvm;
 
+PseudoProbeHandler::~PseudoProbeHandler() = default;
+
 void PseudoProbeHandler::emitPseudoProbe(uint64_t Guid, uint64_t Index,
                                          uint64_t Type, uint64_t Attr,
                                          const DILocation *DebugLoc) {
@@ -35,7 +37,10 @@ void PseudoProbeHandler::emitPseudoProbe(uint64_t Guid, uint64_t Index,
     auto Name = SP->getLinkageName();
     if (Name.empty())
       Name = SP->getName();
-    uint64_t CallerGuid = Function::getGUID(Name);
+    // Use caching to avoid redundant md5 computation for build speed.
+    uint64_t &CallerGuid = NameGuidMap[Name];
+    if (!CallerGuid)
+      CallerGuid = Function::getGUID(Name);
     uint64_t CallerProbeId = PseudoProbeDwarfDiscriminator::extractProbeIndex(
         InlinedAt->getDiscriminator());
     ReversedInlineStack.emplace_back(CallerGuid, CallerProbeId);
diff --git a/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.h b/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.h
index f2026a118bf5..7d5e51218693 100644
--- a/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.h
+++ b/llvm/lib/CodeGen/AsmPrinter/PseudoProbePrinter.h
@@ -26,9 +26,12 @@ class DILocation;
 class PseudoProbeHandler : public AsmPrinterHandler {
   // Target of pseudo probe emission.
   AsmPrinter *Asm;
+  // Name to GUID map, used as caching/memoization for speed.
+  DenseMap<StringRef, uint64_t> NameGuidMap;
 
 public:
   PseudoProbeHandler(AsmPrinter *A) : Asm(A){};
+  ~PseudoProbeHandler() override;
 
   void emitPseudoProbe(uint64_t Guid, uint64_t Index, uint64_t Type,
                        uint64_t Attr, const DILocation *DebugLoc);
diff --git a/llvm/lib/CodeGen/AsmPrinter/WasmException.cpp b/llvm/lib/CodeGen/AsmPrinter/WasmException.cpp
index 352a33e8639d..a17a2ca2790e 100644
--- a/llvm/lib/CodeGen/AsmPrinter/WasmException.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/WasmException.cpp
@@ -18,16 +18,25 @@
 using namespace llvm;
 
 void WasmException::endModule() {
-  // This is the symbol used in 'throw' and 'catch' instruction to denote this
-  // is a C++ exception. This symbol has to be emitted somewhere once in the
-  // module.  Check if the symbol has already been created, i.e., we have at
-  // least one 'throw' or 'catch' instruction in the module, and emit the symbol
-  // only if so.
-  SmallString<60> NameStr;
-  Mangler::getNameWithPrefix(NameStr, "__cpp_exception", Asm->getDataLayout());
-  if (Asm->OutContext.lookupSymbol(NameStr)) {
-    MCSymbol *ExceptionSym = Asm->GetExternalSymbolSymbol("__cpp_exception");
-    Asm->OutStreamer->emitLabel(ExceptionSym);
+  // These are symbols used to throw/catch C++ exceptions and C longjmps. These
+  // symbols have to be emitted somewhere once in the module. Check if each of
+  // the symbols has already been created, i.e., we have at least one 'throw' or
+  // 'catch' instruction with the symbol in the module, and emit the symbol only
+  // if so.
+  //
+  // But in dynamic linking, it is in general not possible to come up with a
+  // module instantiating order in which tag-defining modules are loaded before
+  // the importing modules. So we make them undefined symbols here, define tags
+  // in the JS side, and feed them to each importing module.
+  if (!Asm->isPositionIndependent()) {
+    for (const char *SymName : {"__cpp_exception", "__c_longjmp"}) {
+      SmallString<60> NameStr;
+      Mangler::getNameWithPrefix(NameStr, SymName, Asm->getDataLayout());
+      if (Asm->OutContext.lookupSymbol(NameStr)) {
+        MCSymbol *ExceptionSym = Asm->GetExternalSymbolSymbol(SymName);
+        Asm->OutStreamer->emitLabel(ExceptionSym);
+      }
+    }
   }
 }
 
diff --git a/llvm/lib/CodeGen/AsmPrinter/WinException.cpp b/llvm/lib/CodeGen/AsmPrinter/WinException.cpp
index b30d9cc12abc..ef57031c7294 100644
--- a/llvm/lib/CodeGen/AsmPrinter/WinException.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/WinException.cpp
@@ -43,6 +43,7 @@ WinException::WinException(AsmPrinter *A) : EHStreamer(A) {
   // platforms use an imagerel32 relocation to refer to symbols.
   useImageRel32 = (A->getDataLayout().getPointerSizeInBits() == 64);
   isAArch64 = Asm->TM.getTargetTriple().isAArch64();
+  isThumb = Asm->TM.getTargetTriple().isThumb();
 }
 
 WinException::~WinException() {}
@@ -330,10 +331,12 @@ const MCExpr *WinException::create32bitRef(const GlobalValue *GV) {
 }
 
 const MCExpr *WinException::getLabel(const MCSymbol *Label) {
-  if (isAArch64)
-    return MCSymbolRefExpr::create(Label, MCSymbolRefExpr::VK_COFF_IMGREL32,
-                                   Asm->OutContext);
-  return MCBinaryExpr::createAdd(create32bitRef(Label),
+  return MCSymbolRefExpr::create(Label, MCSymbolRefExpr::VK_COFF_IMGREL32,
+                                 Asm->OutContext);
+}
+
+const MCExpr *WinException::getLabelPlusOne(const MCSymbol *Label) {
+  return MCBinaryExpr::createAdd(getLabel(Label),
                                  MCConstantExpr::create(1, Asm->OutContext),
                                  Asm->OutContext);
 }
@@ -561,8 +564,8 @@ InvokeStateChangeIterator &InvokeStateChangeIterator::scan() {
 ///   struct Table {
 ///     int NumEntries;
 ///     struct Entry {
-///       imagerel32 LabelStart;
-///       imagerel32 LabelEnd;
+///       imagerel32 LabelStart;       // Inclusive
+///       imagerel32 LabelEnd;         // Exclusive
 ///       imagerel32 FilterOrFinally;  // One means catch-all.
 ///       imagerel32 LabelLPad;        // Zero means __finally.
 ///     } Entries[NumEntries];
@@ -664,7 +667,7 @@ void WinException::emitSEHActionsForRange(const WinEHFuncInfo &FuncInfo,
     AddComment("LabelStart");
     OS.emitValue(getLabel(BeginLabel), 4);
     AddComment("LabelEnd");
-    OS.emitValue(getLabel(EndLabel), 4);
+    OS.emitValue(getLabelPlusOne(EndLabel), 4);
     AddComment(UME.IsFinally ? "FinallyFunclet" : UME.Filter ? "FilterFunction"
                                                              : "CatchAll");
     OS.emitValue(FilterOrFinally, 4);
@@ -949,8 +952,15 @@ void WinException::computeIP2StateTable(
       if (!ChangeLabel)
         ChangeLabel = StateChange.PreviousEndLabel;
       // Emit an entry indicating that PCs after 'Label' have this EH state.
+      // NOTE: On ARM architectures, the StateFromIp automatically takes into
+      // account that the return address is after the call instruction (whose EH
+      // state we should be using), but on other platforms we need to +1 to the
+      // label so that we are using the correct EH state.
+      const MCExpr *LabelExpression = (isAArch64 || isThumb)
+                                          ? getLabel(ChangeLabel)
+                                          : getLabelPlusOne(ChangeLabel);
       IPToStateTable.push_back(
-          std::make_pair(getLabel(ChangeLabel), StateChange.NewState));
+          std::make_pair(LabelExpression, StateChange.NewState));
       // FIXME: assert that NewState is between CatchLow and CatchHigh.
     }
   }
diff --git a/llvm/lib/CodeGen/AsmPrinter/WinException.h b/llvm/lib/CodeGen/AsmPrinter/WinException.h
index feea05ba63ad..638589adf0dd 100644
--- a/llvm/lib/CodeGen/AsmPrinter/WinException.h
+++ b/llvm/lib/CodeGen/AsmPrinter/WinException.h
@@ -39,6 +39,9 @@ class LLVM_LIBRARY_VISIBILITY WinException : public EHStreamer {
   /// True if we are generating exception handling on Windows for ARM64.
   bool isAArch64 = false;
 
+  /// True if we are generating exception handling on Windows for ARM (Thumb).
+  bool isThumb = false;
+
   /// Pointer to the current funclet entry BB.
   const MachineBasicBlock *CurrentFuncletEntry = nullptr;
 
@@ -77,6 +80,7 @@ class LLVM_LIBRARY_VISIBILITY WinException : public EHStreamer {
   const MCExpr *create32bitRef(const MCSymbol *Value);
   const MCExpr *create32bitRef(const GlobalValue *GV);
   const MCExpr *getLabel(const MCSymbol *Label);
+  const MCExpr *getLabelPlusOne(const MCSymbol *Label);
   const MCExpr *getOffset(const MCSymbol *OffsetOf, const MCSymbol *OffsetFrom);
   const MCExpr *getOffsetPlusOne(const MCSymbol *OffsetOf,
                                  const MCSymbol *OffsetFrom);